arXiv subjects

Wenyan Xu

Publications and source records attributed to Wenyan Xu.

15 recordsLinked to original sources

arXiv2026

BMFA: Boundary-Minority Free-Energy Adaptive Screening

Vision Transformers process spatially redundant tokens efficiently only when coarse token summaries preserve the evidence required by exponential attention aggregation. We identify a boundary-minority underestimation failure in which a spatially small, high-response region contributes dominant Gibbs mass while remaining nearly invisible to a block mean. We formalize the failure through the discrepancy between normalized log-mean-exp free energy and mean summarization, prove that minority Gibbs mass can remain non-vanishing as its spatial support and mean contribution vanish, and characterize the limitations of finite-order moment corrections. Building on the resulting analysis, we introduce Boundary-Minority Free-Energy Adaptive Screening (BMFA), which constructs a hierarchical piecewise-constant approximation and recursively refines blocks according to a computable lower-bound increment of local free energy. Controlled synthetic tests, COCO and LVIS diagnostic probes, closed-loop DeiT-Tiny evaluations, and ImageNet-1K experiments establish a consistent evidence chain. BMFA reduces the mean synthetic underestimate from 2.582 to 0.261 at a 5.794% leaf ratio, lowers the COCO image-edge mean gap from 2.254 to 0.526, and preserves 71.520% ImageNet Top-1 accuracy at a 55.861% leaf ratio. The current prototype evaluates selection quality after full QK computation; the reported leaf ratio therefore characterizes representation granularity rather than verified sparse-kernel speedup.

cs.CV↗

arXiv2026

Stage-Aware Adaptation and Distribution Calibration for Subject-Driven Personalized Text-to-Image Generation

Subject-driven personalized text-to-image generation requires a pretrained diffusion model to acquire a specific subject from a few reference images while preserving subject identity, following novel text prompts, and maintaining sample diversity. Existing optimization-based methods instantiate subject adaptation through full fine-tuning, textual embedding optimization, or low-rank parameter updates; PaRa further constrains personalization from the perspective of parameter rank reduction. However, a uniform low-rank constraint or a uniform adapter strength cannot explicitly distinguish the capacity requirements of different denoising stages. Moreover, inference-time candidate selection driven mainly by identity similarity may compress the selected samples in the visual representation space. We decompose the problem into two complementary components: SPaRa denotes training-side stage-aware low-rank adaptation, DCAL denotes inference-side distribution-calibrated candidate selection, and SPaRa-DCAL denotes the combined framework. Theoretical analysis shows that timestep-dependent scaling controls the effective perturbation magnitude of a low-rank adapter, while identity-biased candidate selection restricts the radius of selected features around the reference center under explicit conditions. Auditable experiments under the SDXL and DreamBooth 30-subject protocol show that DCAL improves 1-LPIPS, CLIP-I, DINO-I, and CLIP-T on a fixed LoRA candidate pool, while revealing a clear trade-off with CLIP/DINO pairwise diversity and pairwise LPIPS. These results indicate that personalized generation should be evaluated through identity consistency, text alignment, and representation diversity rather than identity metrics alone.

cs.CV↗

arXiv2025

MMM-Fact: A Multimodal, Multi-Domain Fact-Checking Dataset with Multi-Level Retrieval Difficulty

Misinformation and disinformation demand fact checking that goes beyond simple evidence-based reasoning. Existing benchmarks fall short: they are largely single modality (text-only), span short time horizons, use shallow evidence, cover domains unevenly, and often omit full articles -- obscuring models' real-world capability. We present MMM-Fact, a large-scale benchmark of 125,449 fact-checked statements (1995--2025) across multiple domains, each paired with the full fact-check article and multimodal evidence (text, images, videos, tables) from four fact-checking sites and one news outlet. To reflect verification effort, each statement is tagged with a retrieval-difficulty tier -- Basic (1--5 sources), Intermediate (6--10), and Advanced (>10) -- supporting fairness-aware evaluation for multi-step, cross-modal reasoning. The dataset adopts a three-class veracity scheme (true/false/not enough information) and enables tasks in veracity prediction, explainable fact-checking, complex evidence aggregation, and longitudinal analysis. Baselines with mainstream LLMs show MMM-Fact is markedly harder than prior resources, with performance degrading as evidence complexity rises. MMM-Fact offers a realistic, scalable benchmark for transparent, reliable, multimodal fact-checking.

cs.SI↗

arXiv2025

PromptSculptor: Multi-Agent Based Text-to-Image Prompt Optimization

The rapid advancement of generative AI has democratized access to powerful tools such as Text-to-Image models. However, to generate high-quality images, users must still craft detailed prompts specifying scene, style, and context-often through multiple rounds of refinement. We propose PromptSculptor, a novel multi-agent framework that automates this iterative prompt optimization process. Our system decomposes the task into four specialized agents that work collaboratively to transform a short, vague user prompt into a comprehensive, refined prompt. By leveraging Chain-of-Thought reasoning, our framework effectively infers hidden context and enriches scene and background details. To iteratively refine the prompt, a self-evaluation agent aligns the modified prompt with the original input, while a feedback-tuning agent incorporates user feedback for further refinement. Experimental results demonstrate that PromptSculptor significantly enhances output quality and reduces the number of iterations needed for user satisfaction. Moreover, its model-agnostic design allows seamless integration with various T2I models, paving the way for industrial applications.

cs.MA↗

arXiv2025

HRFT: Mining High-Frequency Risk Factor Collections End-to-End via Transformer

In quantitative trading, transforming historical stock data into interpretable, formulaic risk factors enhances the identification of market volatility and risk. Despite recent advancements in neural networks for extracting latent risk factors, these models remain limited to feature extraction and lack explicit, formulaic risk factor designs. By viewing symbolic mathematics as a language where valid mathematical expressions serve as meaningful "sentences" we propose framing the task of mining formulaic risk factors as a language modeling problem. In this paper, we introduce an end to end methodology, Intraday Risk Factor Transformer (IRFT), to directly generate complete formulaic risk factors, including constants. We use a hybrid symbolic numeric vocabulary where symbolic tokens represent operators and stock features, and numeric tokens represent constants. We train a Transformer model on high frequency trading (HFT) datasets to generate risk factors without relying on a predefined skeleton of operators. It determines the general form of the stock volatility law, including constants. We refine the predicted constants using the Broyden Fletcher Goldfarb Shanno (BFGS) algorithm to mitigate non linear issues. Compared to the ten approaches in SRBench, an active benchmark for symbolic regression (SR), IRFT achieves a 30% higher investment return on the HS300 and SP500 datasets, while achieving inference times that are orders of magnitude faster than existing methods in HF risk factor mining tasks.

cs.CE↗

arXiv2025

Mining Intraday Risk Factor Collections via Hierarchical Reinforcement Learning based on Transferred Options

Traditional risk factors like beta, size/value, and momentum often lag behind market dynamics in measuring and predicting stock return volatility. Statistical models like PCA and factor analysis fail to capture hidden nonlinear relationships. Genetic programming (GP) can identify nonlinear factors but often lacks mechanisms for evaluating factor quality, and the resulting formulas are complex. To address these challenges, we propose a Hierarchical Proximal Policy Optimization (HPPO) framework for automated factor generation and evaluation. HPPO uses two PPO models: a high-level policy assigns weights to stock features, and a low-level policy identifies latent nonlinear relationships. The Pearson correlation between generated factors and return volatility serves as the reward signal. Transfer learning pre-trains the high-level policy on large-scale historical data, fine-tuning it with the latest data to adapt to new features and shifts. Experiments show the HPPO-TO algorithm achieves a 25\% excess return in HFT markets across China (CSI 300/800), India (Nifty 100), and the US (S\&P 500). Code and data are available at https://github.com/wencyxu/HRL-HF_risk_factor_set.

cs.CE↗

arXiv2025

FinMultiTime: A Four-Modal Bilingual Dataset for Financial Time-Series Analysis

Pure time series forecasting tasks typically focus exclusively on numerical features; however, real-world financial decision-making demands the comparison and analysis of heterogeneous sources of information. Recent advances in deep learning and large scale language models (LLMs) have made significant strides in capturing sentiment and other qualitative signals, thereby enhancing the accuracy of financial time series predictions. Despite these advances, most existing datasets consist solely of price series and news text, are confined to a single market, and remain limited in scale. In this paper, we introduce FinMultiTime, the first large scale, multimodal financial time series dataset. FinMultiTime temporally aligns four distinct modalities financial news, structured financial tables, K-line technical charts, and stock price time series across both the S&P 500 and HS 300 universes. Covering 5,105 stocks from 2009 to 2025 in the United States and China, the dataset totals 112.6 GB and provides minute-level, daily, and quarterly resolutions, thus capturing short, medium, and long term market signals with high fidelity. Our experiments demonstrate that (1) scale and data quality markedly boost prediction accuracy; (2) multimodal fusion yields moderate gains in Transformer models; and (3) a fully reproducible pipeline enables seamless dataset updates.

cs.CE↗

arXiv2025

Learning Explainable Stock Predictions with Tweets Using Mixture of Experts

Stock price movements are influenced by many factors, and alongside historical price data, tex-tual information is a key source. Public news and social media offer valuable insights into market sentiment and emerging events. These sources are fast-paced, diverse, and significantly impact future stock trends. Recently, LLMs have enhanced financial analysis, but prompt-based methods still have limitations, such as input length restrictions and difficulties in predicting sequences of varying lengths. Additionally, most models rely on dense computational layers, which are resource-intensive. To address these challenges, we propose the FTS- Text-MoE model, which combines numerical data with key summaries from news and tweets using point embeddings, boosting prediction accuracy through the integration of factual textual data. The model uses a Mixture of Experts (MoE) Transformer decoder to process both data types. By activating only a subset of model parameters, it reduces computational costs. Furthermore, the model features multi-resolution prediction heads, enabling flexible forecasting of financial time series at different scales. Experimental results show that FTS-Text-MoE outperforms baseline methods in terms of investment returns and Sharpe ratio, demonstrating its superior accuracy and ability to predict future market trends.

cs.CE↗

arXiv2025

Delay Performance Analysis with Short Packets in Intelligent Machine Network

With the rapid development of delay-sensitive services happened in industrial manufacturing, Internet of Vehicles, and smart logistics, more stringent delay requirements are put forward for the intelligent machine (IM) network. Short packet transmissions are widely adopted to reduce delay in IM networks. However, the delay performance of an IM network has not been sufficiently analyzed. This paper applies queuing theory and stochastic geometry to construct network model and transmission model for downlink communication, respectively, proposes and derives the following three metrics, e.g., the transmission success probability (with delay as the threshold), expected delay, and delay jitter. To accurately characterize the transmission delay with short packets, the finite blocklength capacity is used to measure the channel transmission rate. Simulation results show that the increase of packet length and IM density significantly deteriorates the three metrics. Short packets are needed to improve the three metrics, especially in high IM density scenarios. The outcomes of this paper provide an important theoretical basis for the optimization design and performance improvement of IM networks.

cs.NI↗

arXiv2015

Graphene-Silicon Layered Structures on Single-crystalline Ir(111) Thin Films

Single-crystalline transition metal films are ideal playing fields for the epitaxial growth of graphene and graphene-base materials. Graphene-silicon layered structures were successfully constructed on Ir(111) thin film on Si substrate with an yttria-stabilized zirconia buffer layer via intercalation approach. Such hetero-layered structures are compatible with current Si-based microelectronic technique, showing high promise for applications in future micro- and nano-electronic devices.

cond-mat.mtrl-sci↗

arXiv2014

A cirquent calculus system with clustering and ranking

Cirquent calculus is a new proof-theoretic and semantic approach introduced by G.Japaridze for the needs of his theory of computability logic. The earlier article "From formulas to cirquents in computability logic" by Japaridze generalized the concept of cirquents to the version with what are termed clusterng and ranking, and showed that, through cirquents with clustering and ranking, one can capture, refine and generalize the so called extended IF logic. Japaridze's treatment of extended IF logic, however, was purely semantical, and no deductive system was proposed. The present paper syntactically constructs a cirquent calculus system with clustering and ranking, sound and complete w.r.t. the propositional fragment of cirquent-based semantics. Such a system can be considered not only a conservative extension of classical propositional logic but also, when limited to cirquents with no more than 2 ranks, an axiomatization of purely propositional extended IF logic in its full generality.

cs.LO↗

arXiv2014

A propositional system induced by Japaridze's approach to IF logic

Cirquent calculus is a new proof-theoretic and semantic approach introduced for the needs of computability logic by G.Japaridze, who also showed that, through cirquent calculus, one can capture, refine and generalize independence-friendly (IF) logic. Specifically, the approach allows us to account for independence from propositional connectives in the same spirit as the traditional IF logic accounts for independence from quantifiers. Japaridze's treatment of IF logic, however, was purely semantical, and no deductive system was proposed. The present paper constructs a formal system sound and complete w.r.t. the propositional fragment of Japaridze's cirquent-based semantics for IF logic. Such a system can thus be considered an axiomatization of purely propositional IF logic in its full generality.

cs.LO↗

arXiv2011

The countable versus uncountable branching recurrences in computability logic

This paper introduces a new simplified version of the countable branching recurrence of Computability Logic, proves its equivalence to the old one, and shows that the basic logic induced by it is a proper superset of the basic logic induced by the uncountable branching recurrence. A further result of this paper is showing that the countable branching recurrence is strictly weaker than the uncountable branching recurrence in the sense that the latter logically implies the former but not vice versa.

cs.LO↗

arXiv2011

The parallel versus branching recurrences in computability logic

This paper shows that the basic logic induced by the parallel recurrence of Computability Logic is a proper superset of the basic logic induced by the branching recurrence. The latter is known to be precisely captured by the cirquent calculus system CL15, conjectured by Japaridze to remain sound---but not complete---with parallel recurrence instead of branching recurrence. The present result is obtained by positively verifying that conjecture. A secondary result of the paper is showing that parallel recurrence is strictly weaker than branching recurrence in the sense that, while the latter logically implies the former, vice versa does not hold.

cs.LO↗

arXiv2011

Soundness and completeness of the cirquent calculus system CL6 for computability logic

Computability logic is a formal theory of computability. The earlier article "Introduction to cirquent calculus and abstract resource semantics" by Japaridze proved soundness and completeness for the basic fragment CL5 of computability logic. The present article extends that result to the more expressive cirquent calculus system CL6, which is a conservative extension of both CL5 and classical propositional logic.

cs.LO↗